This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH][4.1]: Fix PR28772: Compile time explosion in the scheduler

Eric Botcazou wrote:
>> In general, compile-time performance regressions are regressions like
>> any other.
> 
> Sure, but this is precisely not one of them. :-)  You can presumably reproduce 
> the problem with all 3.x and 4.x compilers, at least.

I've taken a more expansive view of compile-time performance
regressions: in particular, I consider them all one big issue.  If you
can improve compile times, in a safe and straightforward way (as this
patch does), then it's OK by me for release branches.  That's not meant
as an endorsement of a giant data structure rewrite, but, in general,
compile-time performance tuning is OK for release branches.

The only concern I might have with this patch is that it does limit
optimization and might therefore pessimize code.  However, as it's been
on mainline for a long time, even that's not a concern.

-- 
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713
Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]